last year, the same government agency, the cdc, put that number much lower at 1 in 88. at first this new number sounds like a huge increase. which led us to dig in and dig deeper and look closer at how the cdc arrived at the new number of 1 in 50. it turns out, it was through a phone survey of parents. the cdc didn't actually evaluate any children. they didn't look at a single medical record or any other documentation. they just called up parents and those parents who responded said ultimately -- came out to be 1 in 50. we're not suggesting these parents lied or misled the cdc, but the question is how accurate was this phone survey? and the cdc is, of course, a serious outfit, not prone to throw around fally information. what are the facts? we're joined by sanjay gupta. sanjay, a year ago, everybody was shocked. the prevalence of autism was 1 in 88. now they're saying 1 in 50. has there really been an increase? >> i think in part it depends how you count these cases. i don't think there's been that significant an increase, but this is more reflective of increased awaren